gov.sandia.cognition.learning.algorithm.gradient

Interface ParameterGradientEvaluator<InputType,OutputType,GradientType>

Type Parameters:

InputType - Input type of the Evaluator. For example, Vector.

OutputType - Input type of the Evaluator. For example, Vector.

GradientType - Type of the gradient. For example, Matrix when the parameter and output types are a Vector.

All Superinterfaces:

java.lang.Cloneable, CloneableSerializable, Evaluator<InputType,OutputType>, java.io.Serializable, Vectorizable

All Known Subinterfaces:

GradientDescendable

All Known Implementing Classes:

DifferentiableFeedforwardNeuralNetwork, DifferentiableGeneralizedLinearModel, FactorizationMachine, GradientDescendableApproximator, MultivariateDiscriminant, MultivariateDiscriminantWithBias, PolynomialFunction, ThreeLayerFeedforwardNeuralNetwork

@CodeReview(reviewer="Kevin R. Dixon",
            date="2008-07-23",
            changesNeeded=false,
            comments={"Minor change to javadoc.","Looks fine."})
public interface ParameterGradientEvaluator<InputType,OutputType,GradientType>
extends Evaluator<InputType,OutputType>, Vectorizable

Interface for computing the derivative of the output with respect to the parameters for a given input. The parameters are exposed through a vector on this object, which implements Vectorizable.

Since:

2.0

Author:

Kevin R. Dixon

Method Summary

Modifier and Type	Method and Description
GradientType	computeParameterGradient(InputType input) Computes the derivative of the output with respect to the parameters for a particular input.

Methods inherited from interface gov.sandia.cognition.evaluator.Evaluator

evaluate

Methods inherited from interface gov.sandia.cognition.math.matrix.Vectorizable

clone, convertFromVector, convertToVector

Method Detail

computeParameterGradient

GradientType computeParameterGradient(InputType input)

Computes the derivative of the output with respect to the parameters for a particular input.

Parameters:

input - Input about which to compute the parameter gradient.

Returns:

Change of the parameters with respect to the output.